In-Page Analytics: Why the Google Analytics Site Overlay report does not track all clicks

by Jens Sorensen on July 6, 2011

I recently had a web analytics query from one of our clients specifically around the analysis they were doing for their library pages using the In-Page Analytics/Site overlay report. Being a council and receiving monthly visits in the tens of thousands it was pretty hard to believe that people were interested in some of the library pages but not necessarily going through to the library catalogue pages.

Not all clicks tracking

Links on the right have no clicks

So what’s going on?

The issue lies with the actual content. The three links we see to the right of the image (Search the Library Catalogue, Renew Library Items, and Archives Catalogue) all receive no clicks. These links however go to the sub domain of the site. But why is that an issue? Well, the In-Page Analytics report is quite misleading, it works by tracking the normal JavaScript snippet on your site that transmits the data to Google each time a page is loaded and therefore Google Analytics can conclude that a click was made, but it actually isn’t tracking the “clicks” itself. Because of this, if a click was made to an outbound link it will not track unless you also had the same tracking code on this external site. In this council example, tracking the clicks to the library pages would be entirely possible through the use of cross domain tracking from the main site to the library sub domain (explained below)

However, if like a lot of council websites or any website for that matter which has outbound links to other sources or even banner ads, the In-Page Analytics will not compute the clicks and therefore, although the report can give you a useful visual representation, the insights available from it are minimal, for me at least, making it a function of GA that I do not use.

Furthermore, the In-Page Analytics report is not able to work with the following types of content:

  • JavaScript links
  • Virtual pageviews created with urchinTracker
  • URL redirects
  • URL rewrite filters
  • Links to subdomain pages
  • Frames

And probably the most annoying  and what this report would be useful for is to identify interaction and usability without the need for A/B testing e.g. if one link on a page was more successful than another. However, because we’re not tracking ‘real’ clicks if there is more than one of the same links on a page (even if the anchor text is different) both of these links would give the same click percentage. Again, this is possible if you added an ID to each of the links, but this would be a hassle and not very good from an SEO perspective.

The solution

If visual representation of your data is what you are after tools like ClickTale or Crazy Egg are much more effective at click tracking. However, as I like to keep all my data in one package and don’t worry so much about visual representation I’d rather get the information in Google Analytics through other reports.

So to get the insight into what is happening on the sub domain through Google Analytics, there are a number of things we could do. Firstly, the library sub domain should have GA tracking added (you should be tracking all your sites after all); this would at the very highest level tell you how much the main council site is referring (traffic sources -> referring sites) but not for a specific page. For this you would need sub domain tracking, which basically says to GA the data from this site (sub domain) is part of the main site (make sure you set up filters though to distinguish between the two though).

Alternatively, if you want to track outbound links that are external to your site (e.g. you don’t own the other site). You can call the track outbound link action to allow the distinction of exits from any page. This action is very useful when looking at page exits (something you should be doing). We know there are good and bad exits depending on the stage of the conversion journey, however, without distinguishing whether or not these are due to outbound clicks (usually good), we might think that our site has an overly high exit rate. There are a number of ways to track outbound links; however I think in order not to artificially increase pageviews its best to use the event action.

Previous post:

Next post: